![The Social Yardstick [folded]](http://davidseah.com/_eecontent/imgcache/images/07/466-0728-socialyardstick.jpg)
Last weekend was Barcamp Manchester in Manchester, New Hampshire, which had a pretty good turnout of around 50 people. This year there was a new event, The Make Room, made possible by the donation of several awesome Make Magazine Boxed Set collections by O'Reilly Books. The various books were handed out throughout the day to the people who obviously needed them, and the Box Sets were reserved as prizes for the Make Room contests.
I first entered the Popsicle Stick Bridge Building Competition, but alas: my clever design relied on craftsmanship to implement correctly. I ended up pulling out of the contest, but this left me with leftover parts for the Chindogu Competition.
Chindogu is the Japanese art of inventing items that solve a problem which are burdened with unfortunate and/or embarrassing side effects that, for all practical purposes, render the invention useless. I seem to enjoy making things in this vein: the Procrastinator's Clock and the Gauntlet of Productivity are two such items that I've posted about in the past. Apparently the muse is still with me; I'm very pleased to present the prize-winning chindogu of Barcamp Manchester 2007, an invention I call The Social Yardstick.
INTRODUCING THE SOCIAL YARDSTICK
The Social Yardstick allows you to always maintain an appropriate distance from the people you care about.
Physically, The Yardstick is a measuring device that collapses to fit in your pocket. The prototype here is constructed out of popsicle sticks and packing tape. Each popsicle stick is labeled a listed below:
- Acquaintance -- This is someone you know only slightly.
- Co-Worker -- Someone you work with regularly, but probably not daily. Casual friends too.
- Co-Conspirator -- Someone you are working with, perhaps sharing a hidden agenda or personal goal. You're close, but not too close.
- Best Friend -- Someone you're pretty close with. They're inside your personal space.
- Close Family -- That is, family members that you actually like.
- Sweetie -- You've slept together. Or really want to.
- You -- This is the end of The Yardstick you hold.
The idea for The Social Yardstick was trigged by something that Lokesh Dhakar said at the table I was working at. I was stretching my popsicle sticks (leftover from my failed bridge) in an accordion-like fashion, and he said something about measuring space. I then thought of a time when some friends of mine were having a secret office romance; I'd observed them talking one day standing way too close together to be casual co-workers. I mentioned this to my friend later in case they didn't want to blow their cover. If they had some way of measuring that space between themselves, they could have maintained discretion with ease! And thus, the Social Yardstick idea was born.
The basic principle: your physical distance from someone often telegraphs how "close" you are. Here's a diagram to help illustrate the point:
In case you're wondering: the distances in the diagram are roughly based on what "felt right", though I can see now that it seems roughly based on the width of a person's body.
USING THE YARDSTICK
To use The Social Yardstick, merely unfold its length and stretch between yourself and the person you are standing near. Read the label on the segment that is closest to the other person, and adjust your distance appropriately. Here are some other example uses:
If you are talking to someone you consider an acquaintance and he/she is standing too close, deploy The Yardstick and make sure that your allocation of personal space is not violated.
Conversely, if you're hanging out with your sweetie, use The Yardstick to close the gap: nothing quite says "I Love You" than ensuring you're both inside each other's personal space, short of...well, you get the idea.
You can also use The Social Yardstick as a relationship discovery tool. Say you are in a situation where you are trying to figure out the relationships between two people so you'll have something to talk about at the water cooler. While both parties are standing in place, have the first person hold the stick close to their body and stretch it toward the second person. The label on the stick segment closest to the second person will reveal their actual relationship! You may be surprised by what you find! Just remember: as with the Heisenberg Uncertainty Principle in quantum physics, the act of directly observing your subjects will alter the measurement (e.g. people cough, avert eye contact, and move apart). You may need to apply indirect means of observation over several sessions to arrive at a conclusive reading.
The pre-calibrated distances on The Social Yardstick are based on Urban New England Standards of Personal Space, which is what I'm most familiar with. You may find that you need to adjust for the differing expectations of your locale and culture. For example, people who live in sparsely-populated rural areas probably need a longer yardstick. Likewise, people who work in loud office environments may have to stand closer to just hear each other. And all bets are off when you're in an elevator or picking a urinal in the Men's Restroom.
If you have built your own Social Yardstick or have made something similar in the past, let us know in the comments below! Support Social Unity through Appropriate Personal Spacing!
:-)
NOTE: This article covers the OLD version of Google Analytics at the time of this writing. The current version of Analytics, launched in 2007, is much different.
At the end of October, I moved to a new server and decided to stop using the web traffic logging provided by Mint. There were three reasons behind this decision:
I was spending way too much time checking my stats.
Someone mentioned to me that the size of my WordPress database was surprisingly large. Let me guess...you're running stats? went the comment.
I was reading in the old host's acceptable use policy for databases that stat logging was specifically disallowed. Ooops.
So I dropped Mint and tried surviving on Google Analytics and the free version of Statcounter. That experiment has ended today, and I've decided to reinstall Mint.
Why I Track Stats
Since the vast majority of blog visitors do not leave comments, watching web statistics is one of the primary feedback mechanisms that help me assess what's going on with my writing. The raw numbers of interest to a blogger are page views and unique visitors; it's very satisfying to see the numbers slowly rise over time. When there's a big spike in traffic, that often means that there is something interesting going on elsewhere on the Internet with my name attached to it somehow, and that's good to know.
Also of interest are search terms used to find the site, popular content, and who has linked to me. To a lesser extent, I'm also interested in what outbound links people are clicking on. This gives me a view of how people are using the site, what's popular, what's not, and what's surprising. I once saw a whole bunch of hits from a webmail system from a university, and I later learned that The Printable CEO™ stuff had been making the rounds there. Fascinating!
All three stat trackers I've mentioned rely on Javascript to tickle a webservice on a server to log a visit. The advantage of these services over regular web server log analyzers (Analog, Webalyzer, HTTPAnalyze) is realtime statistics; I can see what's going on at any time of the day to watch trends unfold. With a server log package, the stats are usually available at the end of the day, when it's too late to do anything about a rise in traffic.
Anyway, the javascript-based stat trackers require that there's a webservice for the database to communicate with. For Google Analytics and Statcounter, the servers are under their control. With Mint, you use your website as the server (requirements: a PHP-enabled web server and a MySQL database). Here's how they compare:
Comparing Cost
Mint costs $30/site, and logs everything to your own MySQL database.
Google Analytics is free and doesn't have any limits I know of. UPDATE: It might actually not store data past a month, not sure.
Statcounter is free for up to 100 logged visits at a time. For site that gets more than 50-100 hits a day, you'll need to pay a monthly fee (minimum: $9 for 1000 hits logged, $29 gets you 25K).
I keep the free version of StatCounter enabled because I've always used it, so I can see the the history of web traffic from the day I started blogging.
Comparing Realtime Utility
Mint reports activity to the minute, and is extendable through plugins (called "Peppers") that add additional tracking capabilities. You can get a list of Mint plugins to see what extensions have already been authored.
Statcounter is also up-to-date on a minute-by-minute basis. However, the aforementioned log limit for the free version makes trendwatching over several days difficult without paying at least $29/month. Since Mint itself costs $30 for one-time use, it's a better deal.
Google Analytics is not up-to-the-minute. It seems to take between 4 hours-24 hours to update with the latest data. This is far better than the week it took when the service was hammered by new registrants, but still Mint and Statcounter handily trounce it. Google Analytics is getting faster, but the delay is such that I don't feel it's particularly real-time.
Comparing Quality of Experience
Mint's front end is attractive, robust, and highly useful. The base installation is good for tracking data over a period of time, searches, and referrers. With extensions, you can track downloads, outclicks, and trends, all within the same slickly-designed GUI. Thumbs up! Very Web 2.0! The one thing that is missing is visitor tracking; I know there is a SessionTracker plugin, but it's not as useful as Statcounter's implementation.
Statcounter looks more like an old-school web app, but it provides lots of useful ways of digging through the actual behaviors of your site visitors. In this regard I like it better than Mint, particularly for its visitor tracking capabilities. You can start with a hit from a search result, and then track the trail of page visits to see how people are exploring your site. The SessionTracker pepper for Mint does allow you to see visitor paths, but it doesn't let you track from the very first hit. It is essentially a stand-alone plugin co-existing in the same UI shell, so it can't integrate the data from other parts of Mint (i.e. the keyword search data is is not cross-referenced into the session trail). Statcounter also seems to do a better job of reverse DNS lookup to convert numeric IP numbers to their domain names without making you wait; they probably cache the DNS information. While Mint CAN do the reverse lookup with the XXX Strong Mint plugin, it takes a lot of time and you will have to wait many seconds for each lookup to complete, not to mention that this may bog down your server if you have a large log; this is not the sort of thing you would want to do on a shared server.
Google Analytics feels to me like a hybrid Web 2.0 application, and at first glance provides TONS of useful information and pretty graphs. It has a neat graphical overlay of popular clicks and their position on the page, neat mapping capabilities, flexible date range-based statistical sorting, and a full complement of export capabilities. It also integrates with your AdWords campaigns and gives you all the information you need to track the performance of your content. However, for my needs I don't find it particularly fun to use because the information I need is buried under layers of menus. I can't track session trails through the site in the way I'd like to. It takes several clicks through several menus to get the referrer, keyword search, and popularity results, and I can never remember which is which. Factor in its lag time in updating, and the experience is less than satisfying. With Mint, I can get the big picture of what's going on with my web traffic all on one page. With Google Analytics, I've got to dig through a lot of reports. Plus, Mint just looks better.
In short, Mint does a great job of giving me the information I need to see trends. Statcounter does a better job of showing how people (anonymously, I should mention) are spending their time exploring the site over a period of time. Google Analytics does a lot of that too, but requires too much clicking and page refreshing to build the same picture, with less detail than either Statcounter or Mint. Perhaps all this clicking and raw data is the nature of "web analytics", and I just don't get it.
So Mint is Back Online. Long live Mint!
I'm sure there are tons of other tracking services that I could look into, but these are the three I have the most experience with. For inexpensive, realtime tracking of web traffic, I've found that I needed Mint back. While I had it uninstalled, I found myself going back to Statcounter every 30 minutes to get that "big picture", because Google Analytics just wasn't updating fast enough AND it's cumbersome to get at the stats that I care most about:
- Current traffic levels by the hour and day
- What's popular
- Where links are coming from
- What's being downloaded
- What outgoing links are being clicked
With Google Analytics, getting that information (if it's even available) takes 8-10 clicks in nested menus. With Mint, it's just a matter of visiting the page and looking at it. If I was running actual web marketing campaigns on my site, I'm sure Google Analytics would be far more useful, but as that's not the case it's not for me. If I had the money I would use Statcounter because I like its drill-down features.
Screenshots: Mint
Here's what the Preferences screen looks like. It's all yummy DHTML / AJAX / Web 2.0. Very pretty and functional (in FireFox, anyway).
And here's my main Mint screen. Additional peppers installed (aka plugins): Outclicks, DLoads, Trends, Notification (this sends me an email if some interesting event occurs, like more pageloads in an hour than is normal), SessionTracker, and Sparks.
I should note that I have been tracking data for less than a day, so there aren't a whole lot of stats to look at. But you get the idea.
Another note: I recently dropped DLoads because this was apparently the cause of the massive server spikes that got me kicked off my old server...I think DLoad's method of reading files into PHP to buffer files caused much memory overusage. I switched to Steve Smith's Download Counter as an alternative.
Dave's Assessment: This is a good middle ground between the nutty data overload that is Google Analytics, and the more intimate tracking provided by StatCounter.
» Link to Mint home page. And a bonus link to digital renaissance man Shaun Inman, the creator of Mint.
Screenshots: Google Analytics
Google Analytics is capable of displaying a lot of interesting data and reporting, but it's just overkill for me. There is so much reporting that I've just hilighted a few of the features here:
Here's the main "executive summary" screen for Google Analytics, when you first log-in. Note that there isn't any data for today yet in the system (it's already 830AM here).
The top content on my site, apparently. I had to hunt-and-click through the menus on the left to get here.
The always-interesting "So how did people get to my site?" referring sources page.
I've never seen this page before, but it looks interesting.
Dave's assessment: Should I ever run a web marketing campaign and needed to learn how to use all this data, I would look at this again.
» Link to the Google Analytics home page.
Screenshots: StatCounter
The very Web 1.0 main page of StatCounter. Very straightforward. Since I've been using StatCounter from the time I started blogging, I can see the traffic plotted all the way back by adjusting the date range. It's a pain in the ass to do that in Google Analytics (and it doesn't seem to save data for more than a few months anyway).
I drilled down into the hits for the root page of "davidseah.com". Since I only have 100 hits logged, and am getting about 200-300 hits per hour, I don't see much detailed information. Here you can see, though, that there's one user who has returned 33 times. And he/she uses a Mac! :-)
This is a different user than the one above, but you can get an idea of what StatCounter allows you to do by drilling further down into the stats. You can actually see what a specific user is doing, and where they're coming from.
You may not know their name, but StatCounter is able to trace the browsing habits of a specific user over time, IF you have enough historical data stored in the databases. That comes back down to how many log entries you're paying for. If I 25,000 log entries ($29/month), I'd be able to see quite a bit more history and build a better picture of what people are really doing.
BTW, the blurred-out data in the screenshots is just the IP address and reverse-lookup. It's some cable modem network in the Netherlands. StatCounter doesn't give you the magical ability to look up personal names.
Dave's assessment: StatCounter allows you to track the browsing patterns of individual users easily, which is something that Mint and Google Analytics can't touch. If I was trying to understand the psychology of my visitors to build some kind of mental image of my core audience, I might use something like this. If this kind of tracking bothers you, you might consider blocking cookies from these sites.
» Link to the StatCounter home page.
3 WEEKS LATER...
It's January 4th, and today I was LifeHackered, which took down my site. The helpful folks at my hosting company (FutureQuest) narrowed down the issue to Mint, which was causing server load spikes in excess of 100+. After they disabled it, loads dropped down back to normal.
I thought about this for a bit, and thought that since Mint peppers don't share data, they might be invoking a LOT of extra database and CPU activity for every page load. And they might be doing it inefficiently. I uploaded another instance of Mint (the old one was locked out) and used that to change the configuration to unload everything except the default pepper. Then I asked FutureQuest to re-enable it, which they did cautiously. Since then the load apparently has been OK (Mint hasn't been taken offline again)...2693 page loads in the last hour, and the load has averaged between 1.5 and 3.5. Not great, but not terrible.
Still, this brings up some high traffic considerations when using Mint on a shared server:
Beware of Peppers. They may cause excess CPU utilization, and get you kicked off your shared server when you experience even moderately-high traffic. And yes, I am already cacheing. I see that there's a benchmark flag built into Mint since 1.24, which might be helpful in tracking down which Peppers are the culprit. However, I'm not sure if it benchmarks the actual ping on each pageload. The default installation by itself seems to work fine.
The Fresh Look is Fleeting. Mint is designed to give you an at-a-glance sense of what's going on with your website. When you get a lot of traffic, that freshness is harder to catch. I spend a lot of time looking at unique referrers, and when they're ALL unique and happening several times a minute, you quickly lose 'em. This is probably when using a regular log analysis tool would make more sense, especially if it can tell you WHEN certain pages became popular.
I've been using the various Task Order Up! variations for the past week, and in general I like the idea and the card format. I'm particularly liking the 4x6 card version 5: they're small enough to keep handy and fit in my back pocket, yet big enough to be noticed on my desk; the 3x5 cards tend to get a little lost.
What's interesting: a different kind of job tracking methodology is starting to fall out of this, which is great because I've hit a wall with intermittent projects. I have about 5 or 6 client projects of constantly shifting priorities. Sometimes these projects go dormant for a week or two; one has actually entered deep hibernation due to an internal reorganization.
As I'm not particularly adept at juggling, I've adapted the Task Order Up into something that retains the check-rail compatible qualities, but is more suitable for tasks that are spread over a longer period of time.
Enter the Intermittent Task Tracker.
The Basic Idea
The basic idea behind all the card-based things I've been doing (other than I love cards) is the maintaining of continuity. Compared to tools like the Task Progress Tracker, the Intermittent Task Tracker focuses on maintaining a collection of cards instead of an overall project list. If you like coming at your projects from a top down perspective, you'll probably like the Task Progress Tracker format. If you like thinking more on the details of a task level, then the Intermittent Task Tracker (and the various forms of Task Order Up!) may appeal to you more.
Another way of looking at it: The Task Progress Tracker is great if you like one piece of paper, and maintain a folder of your projects as your continuity-maintaining device. The Intermittent Task Tracker is great if you like multiple cards, and have a recipe box, thumbtack, or check rail system. That's my theory, anyway.
Compared to the Task Order Up: The ITT is designed for use over a period spanning more than a week, when you have projects that don't have hard deadlines. Another project category would be hourly-rate type jobs like maintenance contracts, retained hours, and other "open billed" projects.
Of course, you can use the Intermittent Task Tracker as a regular Task Order Up card...nothing's stopping ya!
Changes from the Task Order Up!
I've dropped the "total hours used" area though because it's not as necessary; while this is useful if you're in a timecard environment and need to sum up your project hours on a weekly basis, that's not the use I need for these.
On the right you can see a scan of an example card; this is pretty much how I'm expecting to use it this week:
There are triangles on the left side of the 15-min bubbles. These mark a new dated entry, so you can separate what hours go with what day with a bit of counting.
Each card can handle up to 20 hours. If you run out of bubbles, just grab a new card and assign a new CARD# at the bottom. This allows you to keep the cards in sequence, and maintain the history of the project in an index card box.
You'll note the use of | marks between bubbles. They're unfortunately a little too close together to easily use (a different color pen would be useful here). These correspond to the / marks in the descriptions; for example, the 5/9 entry shows a "kickoff meeting" and a "first draft". The | mark in the bubbles for the first entry show the split of time: 1hr 15min for the kickoff, and another 30 min for the design draft.
Each description has a dated entry. This is new; the Task Order Up, by comparison, doesn't need these because the tasks are supposed to be short and doable within the span of a week. For intermittent tasks spread over weeks, the date is necessary.
The very last entry shows an example of estimated time. I traced empty circles indicating how much time I thought this task might take to complete, and filled them in as I went. As it so happened, it took less time, and noted this.
Since the job is complete, I wrote closed job at the bottom of the card, just so it's clear it's terminated. This is also indicated by the "1 / 1" at the bottom of the card; you would add the "/ 1" when you know the job is complete. If you run out of space on the card, file it away and make new cards with an incremented card number.
The jobcode is a system I use internally; I assign every incoming prospect a jobcode number, and create a corresponding folder on my production computer's _local_projects folder. This folder is periodically backed up to my main fileserver, or whenever I feel nervous. That number is also used for invoicing, naming BaseCamp projects, and for my SVN repository. One of these days I'll have to automate the whole thing, but I've been too lazy to do it.
As others have noted, you can always write notes on the back of the card!
So that's it! I'll be using this format this week as I close out some lingering projects.
Download the Intermittent Task Tracker
» Download the Intermittent Task Tracker 4x6
» PrintableCEO-ITT01-4X6.pdf
Enjoy!
I feel I must apologize for the sheer number of Task Order Up variations that are cropping up. I have no idea which ones are the most popular...if you have a favorite format or size, let me know so I can keep it alive.
» This article is part of The Printable CEO Series
